Red Rocket - NO OLPF or Noise Reduction

[Frank Cueto]

Is it a "tad noisier" just to you, or a "tad noiser" to every layman?

[ericyoung]

I guess then, it's not a night and day difference, so will come down to a judgement call?

Still very needed for dailies and sameday offline, but useful to know there is some compromise.

[Graeme Nattress]

It's horses for courses - I've seen people with good eyes prefer both. And what I prefer is not what you may prefer, or your footage prefer.

Graeme

[Frank Cueto]

Ok, But wouldn't it be simpler to offer the same debayering across all your debayer engines?

Not trying to be nitpicking. Only trying to understand this to it's fullest.

[roryhinds]

If you're going to Film Out or to DCP I think most people would want CPU processed and avoid the Rocket for finishing

Would you agree?

[Graeme Nattress]

I wouldn't know - I've not surveyed people on the subject. I know what I like in an image, but I also know not everyone agrees with me.

Graeme

[jake blackstone]

Digital Vision recommends using their DVO Sharpen (Digital Vision Optics) even instead of SDK OLPF.
From their FilmMaster user manual:
"Digital Vision DVO Sharpen will often give better results", which I personally found to be the case.
And D_Vision new Clarity NR tool is par excellence.
So, personally, I see no reason not to use RR for finishing work. It all depends on what you use to finish with...

[Peter Moretti]

Graeme,

Is it fair to say that any NR done by RC-X could be replicated by applying an NR filter to RR created files?

If so, is there any type of post NR filter that you would particularly recommend?

Thanks much as always ;).

[Hans von Sonntag]

Hi,

Since early 2008 I never used OPLF compensation or NR within the RedTools. I find that noise reduction, if necessary at all, can be done better/more precisely in the post production process, plus that most projects need a grain/noise management anyway which can not be done inside RedCineX.

The same applies to OLPF compensation which is more or less sharpening done in the debayering process. While the RedOne is comparably soft when compared to other digital cameras it carries so much detail (when focused properly) that enhancing the perceived sharpness should be done later in post. The perceived sharpness depends very much on the lens used (here are big differences to see), the object filmed (a forest vs a baby skin), and of course DOF.

We all know that perception is the measure. There are too many changing factors later to be tackled in post that I don't use OLPF compensation and NR. I just render out the plain footage (RedColor2 and RedLogFilm, thanks Red for this advancement!). IMHO, the discussion around Red's Pro and Cons would be much lesser and better for Red if their would be less options in the footage development process. People want it easy.

Hans

[Uli Plank]

I'd like to second what Graeme wrote before: any kind of filtering you'd like to apply is better done at the latest point in the chain. There are several reasons for this:

– There are specialized software filters for denoising, sharpening and the like which are much better than anything RCX offers, like Neatvideo, but they are slow because of their complexity. So you'd want to use them only on the footage you are really using in the final product.
– Sometimes you'd want to apply such filters only to specific areas of the image, like the darker parts, with the help of masks (luma, in this case). You can't do that in RCX.
– You can't judge the need for advanced filtering before your grading is final, since it might make problems more visible or cover them up.

Finally, if you don't finish in 4K, but in 2K or 1080, it's great to work with the RR for a full decode to 4K (which would take forever in software). You can pan or crop in that image without resolution loss. But when you finally scale it down, a lot of noise will be covered up without any additional filtering. OTOH you may want to add some sharpening after that stage.

Just my two cents (Euro) ;-)